multimedia/dav1d: Update to 1.5.3
Changelog:
Changes for 1.5.3 'Sonic':
--------------------------
1.5.3 is a minor release of dav1d, focused on RISC-V and maintenance:
- Misc small optimizations
- RISC-V assembly optimizations for ipred, emu_edge and w_mask,
and VLEN 512 for blend functions
- Fix issue with ivf files with 0 frames in tools
Changes for 1.5.2 'Sonic':
--------------------------
1.5.2 is a minor release of dav1d, focused on maintenance:
- minor speed improvement in recon
- improvements on loongarch symboles visibility and asm
[4 lines not shown]
mit-krb5: update to 1.22.2
Major changes in 1.22.2 (2026-01-29)
------------------------------------
* Fix a SPNEGO packet parsing bug which could cause GSS mechanism
negotiation failure.
Major changes in 1.22.1 (2025-08-20)
------------------------------------
* Fix a vulnerability in GSS MIC verification [CVE-2025-57736].
Major changes in 1.22 (2025-08-05)
----------------------------------
User experience:
* The libdefaults configuration variable "request_timeout" can be set
to limit the total timeout for KDC requests. When making a KDC
request, the client will now wait indefinitely (or until the request
timeout has elapsed) on a KDC which accepts a TCP connection,
without contacting any additional KDCs. Clients will make fewer DNS
[27 lines not shown]
m4: update to 1.4.21.
* Noteworthy changes in release 1.4.21 (2026-02-06) [stable]
** Fix the `eval' builtin to reject input like `0x' that C does not
accept as an integer literal, rather than silently treating it as zero
(present since "the beginning").
** Fix the `define' and `pushdef' builtins to always warn when attempting
to concatenate the result of using the `defn' macro on a builtin macro
with anything else. This is a counterpart to the change in m4 1.4.11
that issues a warning when `defn' with multiple arguments attempts to
perform concatenation. Attempting to use the output of `defn' anywhere
other than a lone argument during macro definition has never been
portable, but now the outcome is always the remaining text, rather
than being dependent on whether `defn' was used before or after the
other text (present since "the beginning").
** Port to glibc 2.43, which implements functions like strchr as macros
[38 lines not shown]
neomutt: update to 20260406.
2026-04-06 Richard Russon \<rich at flatcap.org\>
* Features
- #4799 Add IMAP connection monitoring, timeouts, and reconnection backoff
* Bug Fixes
- #4755 Fix off-by-one in ARRAY_RESERVE()
- #4819 Fix IMAP crash when changing folders
- Fix freeze in IMAP with OpenSSL
- Fix blocking IMAP pipeline drain with many mailboxes
- Fix crash on threaded index update
- Fix crash on NULL thread in display subject
- Fix menu half-down when data fits in one page
- Fix alias search prompt
- Fix index tag colours
- Protect hooks against unhook during execution
- Fix leak in pattern alias search
- Consistently convert headers to LF EOL convention
* Changed Config
[12 lines not shown]
update to version 9.2.0315.
Changes:
- patch 9.2.0315: missing bound-checks
- patch 9.2.0314: channel: can bind to all network interfaces
- patch 9.2.0313: Callback channel not registered in GUI
- patch 9.2.0312: C-type names are marked as translatable
- patch 9.2.0311: redrawing logic with text properties can be improved
- patch 9.2.0310: unnecessary work in vim_strchr() and find_term_bykeys()
- patch 9.2.0309: Missing out-of-memory check to may_get_cmd_block()
- runtime(log): clean up and modernize log syntax
- patch 9.2.0308: Error message E1547 is wrong
- patch 9.2.0307: more mismatches between return types and documentation
- patch 9.2.0306: runtime(tar): some issues with lz4 support
- runtime(vim9): Fix dist#vim9#Open() spaced paths and SIGPIPE crashes
- patch 9.2.0305: mismatch between return types and documentation
- patch 9.2.0304: tests: test for 9.2.0285 doesn't always fail without the fix
- patch 9.2.0303: tests: zip plugin tests don't check for warning message properly
- patch 9.2.0302: runtime(netrw): RFC2396 decoding double escaping spaces
[117 lines not shown]